TKA Pain: Machine Learning Identification
- A combination of unsupervised and supervised machine learning algorithms may help clinicians identify patients undergoing total knee arthroplasty (TKA) who are likely to experience difficult-to-control pain, according to...
- "By identifying who is at risk for difficult-to-control pain, we will be better able to tailor individualized care for these patients thru targeted patient education and prehabilitation before...
- Chew and colleagues analyzed data from 17,200 patients undergoing TKA from April 2021 to October 2024.
A cutting-edge machine learning algorithm now identifies pain patterns following total knee arthroplasty (TKA), revolutionizing post-operative care. this breakthrough helps tailor pain management,predicting which patients may experience “challenging-to-control pain.” researchers analyzed data from over 17,000 TKA patients,examining pain levels,opioid consumption,and pain progression in the initial 72 hours. The study found specific patient clusters based on pain archetypes, with younger age, higher BMI, and preoperative opioid use correlating with more challenging pain control. News Directory 3 is proud to report on advances shaping modern healthcare.Discover what’s next in personalized pain management strategies.
Machine learning identifies pain patterns after knee replacement
A combination of unsupervised and supervised machine learning algorithms may help clinicians identify patients undergoing total knee arthroplasty (TKA) who are likely to experience difficult-to-control pain, according to a study.
“By identifying who is at risk for difficult-to-control pain, we will be better able to tailor individualized care for these patients thru targeted patient education and prehabilitation before surgery, as well as more optimized pain control after surgery,” said Justin Chew, MD, PhD, regional anesthesiology and acute pain medicine fellow at Hospital for Special Surgery. He presented the data at the 50th Annual Regional Anesthesiology and Acute Pain Medicine Meeting. ”Controlling pain in the immediate postoperative period could prevent acute pain from transitioning into chronic pain, which would ultimately defeat the purpose of the surgery – to relieve pain and improve physical function.”
Methods
Chew and colleagues analyzed data from 17,200 patients undergoing TKA from April 2021 to October 2024.
Outcomes measured included average numeric rating scale (NRS) score,median NRS score,maximum NRS score,normalized area under pain vs. time curve, number of pain scores greater than 4 in the first 24 hours, number of pain scores greater than 7 in the first 24 hours, and morphine milligram equivalents consumption during hospitalization.
Researchers also used an unsupervised learning algorithm to analyze dynamic pain progression up to 72 hours after surgery.
Results
Using the algorithm, chew said his team identified two patient clusters separated by pain archetypes. One cluster had relatively well-controlled postoperative pain,while the other had postoperative pain deemed “difficult to control.”
Chew said the cluster with difficult-to-control pain had higher baseline pain levels after surgery that persisted as nerve blocks wore off and remained difficult to control during thier hospital stay.
In addition, this group had nearly 50% greater opioid consumption postoperatively and nearly double the rate of chronic pain consults required to manage their pain.
Chew also noted a 61% classification accuracy rate with the unsupervised learning model.
“Our classification accuracy rate suggests that our dataset is still missing other predictive factors that are difficult to capture, such as differences in surgical technique,” Chew said. “In the future, we are hoping to continue to explore additional predictive factors that will continue to improve these prediction results.”
For more data:
Justin Chew, MD, PhD, wishes to be contacted through mediarelations@hss.edu.
Sources/Disclosures
Collapse
Chew J, et al. e-Poster 6758. Presented at: Anesthesiology and Acute Pain Medicine Meeting; May 1-3, 2025; Orlando.
Disclosures:
Chew reports no relevant financial disclosures.